Fix S3 bucket name extraction from origin URL

[pradyumna1]

As per AWS docs, bucket name can consist of periods. If an S3 bucket is being used as origin, we need to extract the string before s3 in <bucket-name>.s3.amazonaws.com. I came up with this simple solution as a quick fix, however, we can also use regex for extracting anything before s3.amazonaws.com from the hostname.
For eg. const bucketName = hostname.slice().replace(/(.*)\.s3\.amazonaws\.com/g, '$1')

[dphang]

Seems reasonable if it always starts with s3, maybe just add some simple tests if possible (unless already covered by other tests).

[danielcondemarin]

I'd agree tests for this are needed to avoid regression later.

[danielcondemarin]

#761 looks related. I might jump later in this branch or create a new one to tidy this up.

[codecov]

Codecov Report

Merging #596 (8877ed4) into master (f730097) will increase coverage by 0.04%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #596      +/-   ##
==========================================
+ Coverage   80.14%   80.18%   +0.04%     
==========================================
  Files          57       58       +1     
  Lines        1838     1842       +4     
  Branches      405      405              
==========================================
+ Hits         1473     1477       +4     
  Misses        307      307              
  Partials       58       58              

Impacted Files	Coverage Δ
...ponents/aws-cloudfront/lib/getBucketNameFromUrl.js	100.00% <100.00%> (ø)
...s-components/aws-cloudfront/lib/getOriginConfig.js	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f730097...8877ed4. Read the comment docs.

[dphang]

Closing due to inactivity for some time, if needed please reopen new PR with tests. Thanks!